Multi-target dynamic intelligent production scheduling optimization method and system for discrete manufacturing industry

Abstract

The invention relates to the technical field of industrial production, in particular to a multi-target dynamic intelligent production scheduling optimization method and system for the discrete manufacturing industry, and the method comprises the steps: obtaining production feature data, such as production tasks, equipment, inventory and energy consumption, building a machine learning model through historical data, risks such as delivery delay, inventory shortage and equipment failure in the production process are predicted; constructing a task priority calculation model, comprehensively considering various risk factors, and calculating priority parameters of production tasks; dividing a plurality of optimization objectives for the production task based on the priority parameters, and creating an objective function for each objective to balance production condition conflicts; constructing a production task priority change strategy, and dynamically adjusting a priority parameter and an optimization target according to production feedback; according to the method, data driving and machine learning technologies are fully utilized, intelligent optimization and dynamic adjustment of the discrete manufacturing industry production process are realized, and the production efficiency and flexibility are improved.

Description

Technical Field

[0001] The present invention relates to the field of industrial production technology, and in particular to a multi-objective dynamic intelligent production scheduling optimization method and system for discrete manufacturing industries. Background Art

[0002] In discrete manufacturing, the complexity of production scheduling has greatly increased with the increase in product types and the diversification of manufacturing processes. Traditional scheduling methods usually use static rules or experience-based scheduling strategies, which are difficult to adapt to the complex and changing production environment in real time during the actual production process. This method often ignores the optimization needs of multiple objectives, such as production efficiency, cost control, delivery guarantee, etc., resulting in waste of production system resources, extended production cycles, and reduced customer satisfaction.

[0003] Traditional scheduling methods mainly rely on experience and simple mathematical models. It is difficult to fully obtain and utilize real-time data on key production factors such as production tasks, equipment status, and inventory levels. It is also unable to accurately predict problems such as delivery delays, equipment failures, and inventory shortages that may occur during the production process, resulting in large deviations between production plans and actual execution, and a lack of comprehensive perception and prediction of the production process. In addition, traditional methods usually determine the priority of production tasks based on order delivery dates or subjective experience, ignoring the dependence of production tasks on production resources such as equipment, materials, and energy, as well as potential risk factors, resulting in unreasonable priority division of production tasks and failure to meet actual production needs.

[0004] Discrete manufacturing usually faces multiple optimization goals such as reducing production costs, improving product quality, and shortening production cycles. Traditional scheduling methods often only focus on a single goal and lack the trade-off between the constraints between multiple goals. They cannot achieve the global optimality of the production process. The production optimization goal is single and it is difficult to balance the conflicts between multiple goals. In addition, the production environment of discrete manufacturing is complex and changeable. Factors such as equipment status and material supply change dynamically over time. Traditional static scheduling plans are difficult to adapt to emergencies in the production process and the dynamic changes of optimization goals, resulting in the scheduling plan being out of touch with actual production, unable to adjust and optimize the production process in a timely manner, and the scheduling plan lacks dynamic adjustment capabilities.

[0005] In view of this, the present invention proposes a multi-objective dynamic intelligent production scheduling optimization method and system for discrete manufacturing industries. Summary of the invention

[0006] To achieve the above-mentioned purpose, the present invention provides a multi-objective dynamic intelligent production scheduling optimization method and system for discrete manufacturing industry. The specific technical scheme is as follows: The multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry comprises:

[0007] Acquiring production characteristic data, wherein the production characteristic data includes production task data, production equipment data, inventory data, and energy consumption data;

[0008] Using historical production feature data, we build and train machine learning models for production forecasting, and predict delivery delay data, inventory shortage data, and equipment failure data during production.

[0009] Build a task priority calculation model to calculate production task priority parameters based on delivery delay data, inventory shortage data, and equipment failure data;

[0010] Based on the production task priority parameters, the production tasks are divided into optimization targets, multiple production task optimization targets are constructed, an objective function is created for each production task optimization target, the production condition conflicts in the objective function are balanced, and the production task optimization targets are allocated according to the production task priority parameters;

[0011] Construct a production task priority change strategy, calculate the production task priority change, adjust and calculate the current production task priority parameters based on production feedback, and update the optimization target based on the changes in the production task priority parameters.

[0012] Preferably, obtain production task data, including the production task start time , production task delivery date , the number of production tasks ; Obtain production equipment data, including processing capabilities , startup rate , Failure rate ; Get inventory data, including current inventory , Safety Stock and the demand during the production task ; Obtain energy consumption data, including energy consumption of individual devices And energy consumption price ;in, Indicates the production task number, Indicates the device number. Indicates the material number;

[0013] Obtain historical production feature data, including historical production task data, historical production equipment data, historical energy consumption data, and historical inventory data; obtain the delivery delay time of the corresponding historical production task based on the historical production record data , Equipment failure interval , Equipment failure times and the number of stock shortages and inventory shortage quantity .

[0014] Preferably, a delivery delay prediction model is constructed based on a long short-term memory network LSTM model, wherein the delivery delay prediction model is composed of multiple LSTM layers and a fully connected layer of the LSTM model; the output of the last fully connected layer of the LSTM model is the predicted delivery delay time. ;

[0015] The training steps of the delivery delay prediction model are as follows:

[0016] Convert historical production task data, historical production equipment data, and historical energy consumption data into first time series data, and use the first time series data as input features of the LSTM model;

[0017] According to the historical production record data, the delivery delay time of each historical production task is obtained as a label; the first time series data and the delivery delay time label are combined to form a delivery delay dataset, and the delivery delay dataset is divided into a delivery delay training set, a delivery delay verification set, and a delivery delay test set. The mean square error is used as the loss function, and the Adam optimizer is used to train the model; during the training process, the hyperparameters of the delivery delay prediction model are adjusted according to the performance on the verification set.

[0018] Preferably, a stock shortage prediction model is constructed based on a gated recurrent unit network GRU model, wherein the stock shortage prediction model is composed of multiple GRU layers and a fully connected layer of the GRU model, and the output of the last fully connected layer of the GRU model is subjected to a Sigmoid activation function to obtain the predicted number of stock shortages. , the predicted inventory shortage quantity is obtained through the ReLU activation function ;

[0019] The training steps of the inventory shortage prediction model are as follows:

[0020] Convert historical inventory data, historical production task data, and historical production equipment data into second time series data, and use the second time series data as input features of the GRU model;

[0021] According to the historical production record data, the number of inventory shortages and the shortage quantity of each material are obtained as labels; the second time series data and the labels of the number of inventory shortages and the shortage quantity are combined into an inventory shortage dataset, and the inventory shortage dataset is divided into an inventory shortage training set, an inventory shortage verification set, and an inventory shortage test set. The binary cross entropy loss function is used to optimize the prediction of inventory shortages, the mean square error loss function is used to optimize the prediction of shortage quantity, and the Adam optimizer is used to train the model; during the training process, the hyperparameters of the inventory shortage prediction model are adjusted according to the performance on the inventory shortage verification set.

[0022] Preferably, a device failure prediction model is constructed based on a combination model of a convolutional neural network (CNN) model and a long short-term memory network (LSTM) model;

[0023] The historical production equipment data and fault record data are converted into the third time series data, and the time features of the third time series data are extracted using a one-dimensional convolutional layer; then, the time feature output of the convolutional layer is used as the input of the LSTM layer of the equipment failure prediction model, and the LSTM layer of the equipment failure prediction model is used to capture the long-term dependencies of the third time series data; the output of the LSTM layer of the equipment failure prediction model passes through two parallel fully connected layers, and the Sigmoid activation function is used to output the predicted number of equipment failures , use the ReLU activation function to output the predicted equipment failure interval ;

[0024] The training steps of the equipment failure prediction model are as follows:

[0025] According to the historical production record data, the number of equipment failures and the equipment failure interval time of each device are obtained as labels; the third time series data and the labels of the number of equipment failures and the equipment failure interval time are combined into an equipment failure data set, and the equipment failure data set is divided into an equipment failure training set, an equipment failure verification set and an equipment failure test set. The binary cross entropy loss function is used to optimize the prediction of the number of equipment failures, the mean square error loss function is used to optimize the prediction of the equipment failure interval time, and the Adam optimizer is used to train the model; during the training process, the hyperparameters of the equipment failure prediction model are adjusted according to the performance on the verification set.

[0026] Preferably, a task priority calculation model is constructed to calculate the production task priority parameters according to the production task's delivery delay time, equipment failure interval time, equipment failure times, and inventory shortage times and inventory shortage quantity; for each production task , calculate the risk factor of delivery delay, equipment failure and inventory shortage;

[0027] Calculate the risk factor of delivery delay : ;

[0028] in, is the average of the delivery delay time of all production tasks, is the average quantity of all production tasks;

[0029] Calculate the risk factor of equipment failure : ;

[0030] in, For production tasks The required equipment set, , is the average number of failures of all devices, is the average time between failures of all equipment;

[0031] Calculate the inventory shortage risk factor : ;

[0032] in, For production tasks The required material collection, is the average number of inventory shortages for all materials, For materials The average inventory shortage quantity;

[0033] For each production task , calculate the comprehensive risk factor : ;in, , , is the weight coefficient, satisfying ;

[0034] According to the comprehensive risk factor , calculate the priority parameter of each production task :

[0035] ;

[0036] in, and Indicates the comprehensive risk factor The minimum and maximum values ​​of .

[0037] Preferably, multiple production task optimization objectives are constructed, an objective function is created for each production task optimization objective, and conflicts in production conditions in the objective function are balanced, including: the production task optimization objectives include minimizing production costs, minimizing product defective rates, and minimizing production task time;

[0038] The objective function of the optimization goal of minimizing production cost is: ; The objective function of the optimization goal of minimizing the product defect rate is: ; The objective function of the optimization goal of minimizing the production task time is: ;

[0039] in, For production tasks production costs, is the total number of production tasks;

[0040] Select the optimization model for the production task according to the priority parameter value and set the priority threshold of the optimization objective function and ,in ;

[0041] when When , the optimization goal of minimizing production cost is triggered; when When , the optimization goal of minimizing the product defect rate is triggered; when , triggering the optimization goal of minimizing the production task time.

[0042] Preferably, a production task priority change strategy is constructed to calculate the production task priority change, and according to production feedback, the current production task priority parameters are adjusted and calculated, including:

[0043] Get real-time production feedback data, including actual start time of production tasks , Actual completion time , Product qualification rate , Equipment Utilization and inventory levels ;

[0044] Calculate the actual delivery delay time for each production task ; Calculate the actual mean time between failures for each device and number of failures ; Calculate the actual number of inventory shortages for each material and the actual inventory shortage quantity ; Use actual data to obtain updated delivery delay risk factors , Equipment failure risk factor and inventory shortage risk factor ; Recalculate the updated comprehensive risk factor of each production task and priority parameters ; Dynamically update the optimization target according to the changes in the production task priority parameters;

[0045] When the updated production task priority parameter When , the optimization goal of minimizing production cost is triggered; when When , the optimization goal of minimizing the product defect rate is triggered; when When , the optimization goal of minimizing the production task time is triggered;

[0046] According to the updated optimization target, production scheduling and optimization are re-performed to obtain a new production scheduling strategy; the new production scheduling strategy is converted into scheduling instructions for discrete production characteristics, which are issued to each production link to guide the production process.

[0047] A multi-objective dynamic intelligent production scheduling optimization system for discrete manufacturing industry, which is used to implement the multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry, including: a data acquisition module, a production prediction module, a priority calculation module, a multi-objective optimization module and an optimization update module;

[0048] The data acquisition module is used to acquire production characteristic data, wherein the production characteristic data includes production task data, production equipment data, inventory data and energy consumption data;

[0049] The production forecasting module uses historical production feature data to build and train a machine learning model for production forecasting, and respectively predicts delivery delay data, inventory shortage data, and equipment failure data during production;

[0050] The priority calculation module is used to construct a task priority calculation model and calculate the production task priority parameters according to the delivery delay data, inventory shortage data and equipment failure data;

[0051] The multi-objective optimization module divides the optimization objectives for the production tasks based on the production task priority parameters, constructs multiple production task optimization objectives, creates an objective function for each production task optimization objective, balances the production condition conflicts in the objective function, and allocates the production task optimization objectives according to the production task priority parameters;

[0052] The optimization update module is used to construct a production task priority change strategy, calculate the production task priority change, adjust and calculate the current production task priority parameters based on production feedback, and update the optimization target based on the change of the production task priority parameters.

[0053] Beneficial effects of the present invention: The present invention provides a comprehensive data basis for production forecasting and optimization by acquiring production tasks, equipment, inventory and energy consumption data, which helps to gain an in-depth understanding of all aspects of the production process and lays the foundation for subsequent analysis and decision-making.

[0054] By training machine learning models to predict delivery delays, inventory shortages and equipment failures, the present invention can discover potential risks and problems in production in advance, provide a basis for formulating response strategies, reduce uncertainties in the production process, and improve production efficiency and product quality.

[0055] The present invention calculates the priority parameters of production tasks by comprehensively considering factors such as delivery delays, inventory shortages and equipment failures, can reasonably evaluate the importance and urgency of tasks, provide a decision-making basis for production scheduling and resource allocation, and help improve production efficiency and customer satisfaction.

[0056] The present invention constructs multiple optimization objectives and creates corresponding objective functions according to task priority parameters, which can flexibly optimize according to different production needs and conditions, balance factors such as production cost, product quality and production time, and achieve comprehensive optimization and improvement of the production process.

[0057] The present invention can adapt to changes and emergencies in the production process by calculating task priority changes in real time and dynamically adjusting according to production feedback, timely update optimization targets, realize dynamic optimization of production scheduling, improve the flexibility and adaptability of the production system, and ensure stable and efficient production. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A flow chart of the multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry provided by the present invention;

[0059] Figure 2 This is a structural diagram of the multi-objective dynamic intelligent production scheduling optimization system for discrete manufacturing industries provided by the present invention. DETAILED DESCRIPTION

[0060] In order to better understand the present invention, a more detailed description will be made of various aspects of the present invention with reference to the accompanying drawings. It should be understood that these detailed descriptions are only descriptions of exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention in any way. Throughout the specification, the same reference numerals refer to the same elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0061] In the accompanying drawings, the size, dimensions and shapes of the elements have been slightly adjusted for ease of illustration. The accompanying drawings are for illustration only and are not strictly drawn to scale. As used herein, the terms "substantially", "approximately" and similar terms are used as terms of approximation, not as terms of degree, and are intended to illustrate the inherent deviations in measurements or calculations that will be recognized by those of ordinary skill in the art. In addition, in the present invention, the order in which the steps are described does not necessarily represent the order in which these processes occur in actual operation, unless otherwise specified or can be derived from the context.

[0062] It should also be understood that expressions such as "comprises", "including", "having", "includes" and / or "comprising" are open rather than closed expressions in this specification, which indicate the presence of the stated features, elements and / or components, but do not exclude the presence of one or more other features, elements, components and / or combinations thereof. In addition, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features rather than just the individual elements in the list. In addition, when describing embodiments of the present invention, "may" is used to mean "one or more embodiments of the present invention". And, the term "exemplary" is intended to refer to an example or illustration.

[0063] Unless otherwise defined, all terms (including engineering terms and scientific and technological terms) used in this document have the same meaning as commonly understood by ordinary technicians in the field to which the present invention belongs. It should also be understood that unless otherwise clearly stated in the present invention, words defined in commonly used dictionaries should be interpreted as having the same meaning as their meaning in the context of the relevant technology, and should not be interpreted in an idealized or overly formal sense.

[0064] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0065] Example 1

[0066] Reference Figure 1 , which is the first embodiment of the present invention, provides a multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry.

[0067] S1: Acquire production characteristic data, wherein the production characteristic data includes production task data, production equipment data, inventory data and energy consumption data.

[0068] Get production task data, including production task start time , production task delivery date , the number of production tasks ; Obtain production equipment data, including processing capabilities , startup rate , Failure rate ; Get inventory data, including current inventory , Safety Stock and the demand during the production task ; Obtain energy consumption data, including energy consumption of individual devices And energy consumption price ;in, Indicates the production task number, Indicates the device number. Indicates the material number.

[0069] Obtain historical production feature data, including historical production task data, historical production equipment data, historical energy consumption data, and historical inventory data; obtain the delivery delay time of the corresponding historical production task based on the historical production record data , Equipment failure interval , Equipment failure times and the number of stock shortages and inventory shortage quantity .

[0070] By acquiring production characteristic data and historical production characteristic data, the necessary data foundation is provided for subsequent production prediction and optimization.

[0071] S1 provides the necessary data foundation for subsequent production prediction and optimization by acquiring production feature data and historical production feature data. These data include information on production tasks, equipment, inventory, energy consumption, and other aspects, as well as data on delivery delays, equipment failures, and inventory shortages in historical production, laying the foundation for analyzing potential risks in the production process and optimizing production scheduling.

[0072] S2: Use historical production feature data to build and train machine learning models for production forecasting, and predict delivery delay data, inventory shortage data, and equipment failure data during production.

[0073] A delivery delay prediction model is constructed based on the long short-term memory network LSTM model. The delivery delay prediction model consists of multiple LSTM layers and a fully connected layer of the LSTM model; the output of the last fully connected layer of the LSTM model is the predicted delivery delay time. .

[0074] The training steps of the delivery delay prediction model are as follows:

[0075] Convert historical production task data, historical production equipment data, and historical energy consumption data into first time series data, and use the first time series data as input features of the LSTM model;

[0076] According to the historical production record data, the delivery delay time of each historical production task is obtained as a label; the first time series data and the delivery delay time label are combined to form a delivery delay dataset, and the delivery delay dataset is divided into a delivery delay training set, a delivery delay verification set and a delivery delay test set. The mean square error (MSE) is used as the loss function, and the Adam optimizer is used to train the model; during the training process, the hyperparameters of the delivery delay prediction model are adjusted according to the performance on the verification set.

[0077] Input Data , the output data is the predicted delivery delay time ;

[0078] A stock shortage prediction model is constructed based on a gated recurrent unit network GRU model. The stock shortage prediction model consists of multiple GRU layers and a fully connected layer of the GRU model. The output of the last fully connected layer of the GRU model is subjected to a Sigmoid activation function to obtain the predicted number of stock shortages. , the predicted inventory shortage quantity is obtained through the ReLU activation function .

[0079] The training steps of the inventory shortage prediction model are as follows:

[0080] Convert historical inventory data, historical production task data, and historical production equipment data into second time series data, and use the second time series data as input features of the GRU model;

[0081] According to the historical production record data, the number of inventory shortages and the shortage quantity of each material are obtained as labels; the second time series data and the labels of the number of inventory shortages and the shortage quantity are combined into an inventory shortage dataset, and the inventory shortage dataset is divided into an inventory shortage training set, an inventory shortage verification set and an inventory shortage test set. The binary cross entropy loss function is used to optimize the inventory shortage prediction, the mean square error loss function is used to optimize the shortage quantity prediction, and the Adam optimizer is used to train the model; during the training process, the hyperparameters of the inventory shortage prediction model are adjusted according to the performance on the inventory shortage verification set.

[0082] The equipment failure prediction model is constructed based on the combined model of the convolutional neural network (CNN) model and the long short-term memory (LSTM) network model.

[0083] The historical production equipment data and fault record data are converted into the third time series data, and the time features of the third time series data are extracted using a one-dimensional convolutional layer; then, the time feature output of the convolutional layer is used as the input of the LSTM layer of the equipment failure prediction model, and the LSTM layer of the equipment failure prediction model is used to capture the long-term dependencies of the third time series data; the output of the LSTM layer of the equipment failure prediction model passes through two parallel fully connected layers, and the Sigmoid activation function is used to output the predicted number of equipment failures , use the ReLU activation function to output the predicted equipment failure interval .

[0084] The training steps of the equipment failure prediction model are as follows:

[0085] According to the historical production record data, the number of equipment failures and the equipment failure interval time of each device are obtained as labels; the third time series data and the labels of the number of equipment failures and the equipment failure interval time are combined into an equipment failure data set, and the equipment failure data set is divided into an equipment failure training set, an equipment failure verification set and an equipment failure test set. The binary cross entropy loss function is used to optimize the prediction of the number of equipment failures, the mean square error loss function is used to optimize the prediction of the equipment failure interval time, and the Adam optimizer is used to train the model; during the training process, the hyperparameters of the equipment failure prediction model are adjusted according to the performance on the verification set.

[0086] Machine learning models are used to predict key indicators in the production process, including delivery delays, inventory shortages, and equipment failures. The accuracy and reliability of predictions are improved through the training and optimization of models such as LSTM, GRU, and CNN. These prediction results provide important decision-making basis for the calculation of subsequent task priorities and production scheduling optimization.

[0087] S3: Build a task priority calculation model to calculate production task priority parameters based on delivery delay data, inventory shortage data, and equipment failure data.

[0088] Construct a task priority calculation model to calculate the production task priority parameters based on the production task's delivery delay time, equipment failure interval time, equipment failure times, inventory shortage times and inventory shortage quantity; for each production task , calculate the risk factor of delivery delay, equipment failure and inventory shortage.

[0089] Calculate the risk factor of delivery delay : ;in, is the average of the delivery delay time of all production tasks, is the average quantity of all production tasks.

[0090] Calculate the risk factor of equipment failure : ;in, For production tasks The required equipment set, , is the average number of failures of all devices, It is the average time between failures of all devices.

[0091] Calculate the inventory shortage risk factor : ;in, For production tasks The required material collection, is the average number of inventory shortages for all materials, For materials The average inventory shortage quantity.

[0092] For each production task , calculate the comprehensive risk factor : ;in, , , is the weight coefficient, satisfying , which can be adjusted according to actual production conditions.

[0093] According to the comprehensive risk factor , calculate the priority parameter of each production task :

[0094] ;

[0095] in, and Represents the comprehensive risk factor The minimum and maximum values ​​of the priority parameter The larger the value, the more production tasks The higher the overall risk, the higher the priority should be given.

[0096] Build a task priority calculation model, comprehensively consider risk factors such as delivery delays, equipment failures, and inventory shortages, and quantify the priority parameters of each production task. By introducing risk factors and weights, the priority can be dynamically adjusted according to the actual production situation. The calculation of task priority parameters provides a clear task importance measurement standard for subsequent production scheduling optimization, which helps to reasonably allocate production resources.

[0097] S4: Based on the production task priority parameters, the optimization objectives are divided into production tasks, multiple production task optimization objectives are constructed, an objective function is created for each production task optimization objective, the production condition conflicts in the objective function are balanced, and the production task optimization objectives are allocated according to the production task priority parameters.

[0098] Construct multiple production task optimization objectives, create an objective function for each production task optimization objective, and balance the production condition conflicts in the objective function, including: the production task optimization objectives include minimizing production costs, minimizing product defective rates, and minimizing production task time.

[0099] The objective function of the optimization goal of minimizing production cost is: ;

[0100] The constraints for minimizing production cost are as follows:

[0101] ;

[0102] ;

[0103] ;

[0104] ;

[0105] The objective function of the optimization goal of minimizing the product defect rate is: .

[0106] Minimize the product defect rate constraint:

[0107] ;

[0108] ;

[0109] ;

[0110] The objective function of the optimization goal of minimizing the production task time is: .

[0111] The constraints for minimizing the production task time are as follows:

[0112] ;

[0113] ;

[0114] ;

[0115] in, For production tasks production costs, is the total number of production tasks.

[0116] Select the optimization model of the production task according to the priority parameter value and set the priority threshold of the optimization objective function. , ,in .

[0117] when When , the optimization goal of minimizing production cost is triggered; when When , the optimization goal of minimizing the product defect rate is triggered; when , triggering the optimization goal of minimizing the production task time.

[0118] By building an optimization function, according to the priority parameters of the production task , a trade-off is made between minimizing production costs, minimizing product defect rates, and minimizing production task time; the constraints of these three optimization functions ensure that the completion of production tasks does not exceed the production capacity of the equipment and meets inventory requirements.

[0119] S5: Construct a production task priority change strategy, calculate the production task priority change, adjust and calculate the current production task priority parameters based on production feedback, and update the optimization target based on the change of the production task priority parameters.

[0120] Get real-time production feedback data, including actual start time of production tasks , Actual completion time , Product qualification rate , Equipment Utilization and inventory levels .

[0121] Calculate the actual delivery delay time for each production task : .

[0122] Calculate the actual time between failures for each device and number of failures :

[0123] ;

[0124] in, For use of equipment The production task set, For production tasks Use equipment The actual number of failures during the period.

[0125] Calculate actual inventory shortages for each material and the actual inventory shortage quantity :

[0126] ;

[0127] in, For production tasks Period Materials The demand, is the indicator function, when The function value is 1 when , otherwise it is 0. The function is used to determine whether there is a shortage of materials required for the production task.

[0128] Use actual data to obtain updated delivery delay risk factors , Equipment failure risk factor and inventory shortage risk factor :

[0129] ;

[0130] ;

[0131] ;

[0132] in, It represents the average number of days actually delayed for all production tasks; is the average value of the actual number of production tasks; Indicates the device The number of actual failures that occurred during the execution of production tasks;

[0133] Indicates material The average number of actual shortages that occurred during the execution of all production tasks; Indicates material The average value of the actual shortage quantity during the execution of the production task.

[0134] Recalculate the updated comprehensive risk factor for each production task and priority parameters :

[0135] ;

[0136] ;

[0137] in, and Represents the updated comprehensive risk factor The minimum and maximum values ​​of .

[0138] Dynamically update optimization objectives based on changes in production task priority parameters.

[0139] When the updated production task priority parameter When , the optimization goal of minimizing production cost is triggered; when When , the optimization goal of minimizing the product defect rate is triggered; when , triggering the optimization goal of minimizing the production task time.

[0140] According to the updated optimization target, production scheduling and optimization are re-performed to obtain a new production scheduling strategy; the new production scheduling strategy is converted into scheduling instructions for discrete production characteristics, which are issued to each production link to guide the production process.

[0141] Build a production task priority change strategy, dynamically adjust task priority parameters based on real-time production feedback data, and update the optimization target accordingly.

[0142] By real-time monitoring of delivery delays, equipment failures, inventory shortages, etc. during the production process, optimization strategies can be adjusted in a timely manner to achieve dynamic optimization of production scheduling; converting optimized production scheduling strategies into discrete scheduling instructions can effectively guide the production process and improve production flexibility and adaptability.

[0143] Example 2

[0144] Reference Figure 2 , which is the second embodiment of the present invention, provides a multi-objective dynamic intelligent production scheduling optimization system for discrete manufacturing industries.

[0145] The system includes: a data acquisition module, a production prediction module, a priority calculation module, a multi-objective optimization module and an optimization update module.

[0146] The data acquisition module is used to acquire production characteristic data, and the production characteristic data includes production task data, production equipment data, inventory data and energy consumption data.

[0147] The production forecasting module uses historical production feature data to build and train a machine learning model for production forecasting, and respectively predicts delivery delay data, inventory shortage data, and equipment failure data during production.

[0148] The priority calculation module is used to construct a task priority calculation model and calculate production task priority parameters based on delivery delay data, inventory shortage data, and equipment failure data.

[0149] The multi-objective optimization module divides optimization objectives for production tasks based on production task priority parameters, constructs multiple production task optimization objectives, creates an objective function for each production task optimization objective, balances production condition conflicts in the objective function, and allocates production task optimization objectives according to production task priority parameters.

[0150] The optimization update module is used to construct a production task priority change strategy, calculate the production task priority change, adjust and calculate the current production task priority parameters based on production feedback, and update the optimization target based on the change of the production task priority parameters.

[0151] Example 3

[0152] The present invention also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable codes, and when the computer-readable codes are executed by the one or more processors, the multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industries as described above may be executed.

[0153] The method or system according to the embodiment of the present invention may also be implemented with the aid of the architecture of the electronic device of the present invention.

[0154] An electronic device may include a bus, one or more CPUs, read-only memory (ROM), random access memory (RAM), a communication port for connecting to a network, input / output components, a hard disk, and the like.

[0155] A storage device in an electronic device, such as a ROM or a hard disk, can store the multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industries provided by the present invention.

[0156] A multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industries includes: obtaining production feature data, wherein the production feature data includes production task data, production equipment data, inventory data, and energy consumption data; using historical production feature data to construct and train a machine learning model for production prediction, and respectively predicting delivery delay data, inventory shortage data, and equipment failure data during production; constructing a task priority calculation model, and calculating production task priority parameters according to delivery delay data, inventory shortage data, and equipment failure data; dividing optimization targets for production tasks based on production task priority parameters, constructing multiple production task optimization targets, creating an objective function for each production task optimization target, balancing production condition conflicts in the objective function, and allocating production task optimization targets according to production task priority parameters; constructing a production task priority change strategy, calculating production task priority changes, adjusting and calculating current production task priority parameters according to production feedback, and updating the optimization targets according to changes in production task priority parameters.

[0157] Furthermore, the electronic device may also include a user interface. Of course, the architecture disclosed in the present invention is only exemplary, and when implementing different devices, one or more components in the electronic device disclosed in the present invention may be omitted according to actual needs.

[0158] Example 4

[0159] The invention discloses a computer readable storage medium.

[0160] The computer-readable storage medium has computer-readable instructions stored thereon.

[0161] When the computer-readable instructions are executed by a processor, the multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industries according to the embodiments of the present invention described with reference to the above drawings can be executed.

[0162] The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. In addition, according to an embodiment of the present invention, the process described above with reference to the flowchart may be implemented as a computer software program.

[0163] For example, the present invention provides a non-temporary machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided by the present invention, for example: obtaining production feature data, the production feature data including production task data, production equipment data, inventory data and energy consumption data; using historical production feature data to build and train a machine learning model for production forecasting, and respectively predicting delivery delay data, inventory shortage data and equipment failure data during production; building a task priority calculation model, and calculating the production task priority parameters according to the delivery delay data, inventory shortage data and equipment failure data; dividing the production tasks into optimization targets based on the production task priority parameters, constructing multiple production task optimization targets, creating an objective function for each production task optimization target, balancing the production condition conflicts in the objective function, and allocating production task optimization targets according to the production task priority parameters; building a production task priority change strategy, calculating the production task priority change, adjusting and calculating the current production task priority parameters according to production feedback, and updating the optimization targets according to the changes in the production task priority parameters.

[0164] When the computer program is executed by the central processing unit (CPU), the above functions defined in the method of the present invention are performed. The method, apparatus, and device of the present invention may be implemented in many ways. For example, the method, apparatus, and device of the present invention may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware.

[0165] The above sequence for the steps of the method is for illustration only, and the steps of the method of the present invention are not limited to the sequence specifically described above unless otherwise specifically stated.

[0166] In addition, in some embodiments, the present invention can also be implemented as a program recorded in a recording medium, which includes machine-readable instructions for implementing the method according to the present invention. Therefore, the present invention also covers a recording medium storing a program for executing the method according to the present invention.

[0167] In addition, the parts of the above technical solutions provided in the embodiments of the present invention that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0168] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry, characterized by: include: Acquiring production characteristic data, wherein the production characteristic data includes production task data, production equipment data, inventory data, and energy consumption data; Using historical production feature data, we build and train machine learning models for production forecasting, and predict delivery delay data, inventory shortage data, and equipment failure data during production. Build a task priority calculation model to calculate production task priority parameters based on delivery delay data, inventory shortage data, and equipment failure data; Based on the production task priority parameters, the production tasks are divided into optimization targets, multiple production task optimization targets are constructed, an objective function is created for each production task optimization target, the production condition conflicts in the objective function are balanced, and the production task optimization targets are allocated according to the production task priority parameters; Construct a production task priority change strategy, calculate the production task priority change, adjust and calculate the current production task priority parameters based on production feedback, and update the optimization target based on the changes in the production task priority parameters.

2. The multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry according to claim 1 is characterized in that: Get production task data, including production task start time , production task delivery date , the number of production tasks ; Obtain production equipment data, including processing capabilities , startup rate , Failure rate ; Get inventory data, including current inventory levels , Safety Stock and the demand during the production task ; Obtain energy consumption data, including energy consumption of individual devices And energy consumption price ;in, Indicates the production task number, Indicates the device number. Indicates the material number; Obtain historical production feature data, including historical production task data, historical production equipment data, historical energy consumption data, and historical inventory data; obtain the delivery delay time of the corresponding historical production task based on the historical production record data , Equipment failure interval , Equipment failure times and the number of stock shortages and inventory shortage quantity .

3. The multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry according to claim 2 is characterized in that: A delivery delay prediction model is constructed based on the long short-term memory network LSTM model. The delivery delay prediction model consists of multiple LSTM layers and a fully connected layer of the LSTM model; the output of the last fully connected layer of the LSTM model is the predicted delivery delay time. ; The training steps of the delivery delay prediction model are as follows: Convert historical production task data, historical production equipment data, and historical energy consumption data into first time series data, and use the first time series data as input features of the LSTM model; According to the historical production record data, the delivery delay time of each historical production task is obtained as a label; the first time series data and the delivery delay time label are combined to form a delivery delay dataset, and the delivery delay dataset is divided into a delivery delay training set, a delivery delay verification set, and a delivery delay test set. The mean square error is used as the loss function, and the Adam optimizer is used to train the model; during the training process, the hyperparameters of the delivery delay prediction model are adjusted according to the performance on the verification set.

4. The multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry according to claim 3 is characterized in that: A stock shortage prediction model is constructed based on a gated recurrent unit network GRU model. The stock shortage prediction model consists of multiple GRU layers and a fully connected layer of the GRU model. The output of the last fully connected layer of the GRU model is subjected to a Sigmoid activation function to obtain the predicted number of stock shortages. , after the ReLU activation function, the predicted inventory shortage quantity is obtained ; The training steps of the inventory shortage prediction model are as follows: Convert historical inventory data, historical production task data, and historical production equipment data into second time series data, and use the second time series data as input features of the GRU model; According to the historical production record data, the number of inventory shortages and the shortage quantity of each material are obtained as labels; The second time series data is combined with the labels of the number of inventory shortages and the number of shortages to form an inventory shortage dataset. The inventory shortage dataset is divided into an inventory shortage training set, an inventory shortage validation set, and an inventory shortage test set. The binary cross entropy loss function is used to optimize the inventory shortage prediction, the mean square error loss function is used to optimize the shortage quantity prediction, and the Adam optimizer is used to train the model. During the training process, the hyperparameters of the stock outage prediction model are adjusted based on the performance on the stock outage validation set.

5. The multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry according to claim 4 is characterized in that: Building an equipment failure prediction model based on a combination of a convolutional neural network (CNN) model and a long short-term memory (LSTM) model; The historical production equipment data and fault record data are converted into the third time series data, and the time features of the third time series data are extracted using a one-dimensional convolutional layer; then, the time feature output of the convolutional layer is used as the input of the LSTM layer of the equipment failure prediction model, and the LSTM layer of the equipment failure prediction model is used to capture the long-term dependencies of the third time series data; the output of the LSTM layer of the equipment failure prediction model passes through two parallel fully connected layers, and the Sigmoid activation function is used to output the predicted number of equipment failures , use the ReLU activation function to output the predicted equipment failure interval ; The training steps of the equipment failure prediction model are as follows: According to the historical production record data, the number of equipment failures and the equipment failure interval time of each equipment are obtained as labels; The third time series data and the labels of the number of equipment failures and the equipment failure intervals are combined into a equipment failure dataset. The equipment failure dataset is divided into an equipment failure training set, an equipment failure verification set, and an equipment failure test set. The binary cross entropy loss function is used to optimize the prediction of the number of equipment failures, and the mean square error loss function is used to optimize the prediction of the equipment failure interval. The Adam optimizer is used to train the model. During the training process, the hyperparameters of the equipment failure prediction model are adjusted according to the performance on the verification set.

6. The multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry according to claim 5 is characterized in that: Construct a task priority calculation model to calculate the production task priority parameters based on the production task's delivery delay time, equipment failure interval time, equipment failure times, inventory shortage times and inventory shortage quantity; for each production task , calculate the risk factor of delivery delay, equipment failure and inventory shortage; Calculate the risk factor of delivery delay : ; in, is the average of the delivery delay time of all production tasks, is the average quantity of all production tasks; Calculate the risk factor of equipment failure : ; in, For production tasks The required equipment set, , is the average number of failures of all devices, is the average time between failures of all equipment; Calculate the inventory shortage risk factor : ; in, For production tasks The required material collection, is the average number of inventory shortages for all materials, For materials The average inventory shortage quantity; For each production task , calculate the comprehensive risk factor : ;in, , , is the weight coefficient, satisfying ; According to the comprehensive risk factor , calculate the priority parameter of each production task : ； in, and Indicates the comprehensive risk factor The minimum and maximum values ​​of .

7. The multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry according to claim 6 is characterized in that: Constructing multiple production task optimization objectives, creating an objective function for each production task optimization objective, and balancing the production condition conflicts in the objective function, including: the production task optimization objectives include minimizing production costs, minimizing product defect rates, and minimizing production task time; The objective function of the optimization goal of minimizing production cost is: ; The objective function of the optimization goal of minimizing the product defect rate is: ; The objective function of the optimization goal of minimizing the production task time is: ; in, For production tasks production costs, is the total number of production tasks; Select the optimization model for the production task according to the priority parameter value and set the priority threshold of the optimization objective function and ,in ; when When , the optimization goal of minimizing production cost is triggered; when When , the optimization goal of minimizing the product defect rate is triggered; when , triggering the optimization goal of minimizing the production task time.

8. The multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry according to claim 7 is characterized in that: Build a production task priority change strategy, calculate the production task priority change, and adjust the current production task priority parameters based on production feedback, including: Get real-time production feedback data, including actual start time of production tasks , Actual completion time , Product qualification rate , Equipment Utilization and inventory levels ; Calculate the actual delivery delay time for each production task ; Calculate the actual mean time between failures for each device and number of failures ; Calculate the actual number of inventory shortages for each material and the actual inventory shortage quantity ; Use actual data to obtain updated delivery delay risk factors , Equipment failure risk factor and inventory shortage risk factor ; Recalculate the updated comprehensive risk factor of each production task and priority parameters ; Dynamically update the optimization target according to the changes in the production task priority parameters; When the updated production task priority parameter When , the optimization goal of minimizing production cost is triggered; when When , the optimization goal of minimizing the product defect rate is triggered; when When , the optimization goal of minimizing the production task time is triggered; According to the updated optimization target, production scheduling and optimization are re-performed to obtain a new production scheduling strategy; the new production scheduling strategy is converted into scheduling instructions for discrete production characteristics, which are issued to each production link to guide the production process.

9. A multi-objective dynamic intelligent production scheduling optimization system for discrete manufacturing industry, which is used to implement the multi-objective dynamic intelligent production scheduling optimization method for discrete manufacturing industry according to any one of claims 1 to 8, characterized in that: include: Data acquisition module, production prediction module, priority calculation module, multi-objective optimization module and optimization update module; The data acquisition module is used to acquire production characteristic data, wherein the production characteristic data includes production task data, production equipment data, inventory data and energy consumption data; The production forecasting module uses historical production feature data to build and train a machine learning model for production forecasting, and respectively predicts delivery delay data, inventory shortage data, and equipment failure data during production; The priority calculation module is used to construct a task priority calculation model and calculate the production task priority parameters according to the delivery delay data, inventory shortage data and equipment failure data; The multi-objective optimization module divides the optimization objectives for the production tasks based on the production task priority parameters, constructs multiple production task optimization objectives, creates an objective function for each production task optimization objective, balances the production condition conflicts in the objective function, and allocates the production task optimization objectives according to the production task priority parameters; The optimization update module is used to construct a production task priority change strategy, calculate the production task priority change, adjust and calculate the current production task priority parameters based on production feedback, and update the optimization target based on the change of the production task priority parameters.

Images